<p>Accurate and efficient control of structural response during seismic events remains a critical challenge in structural engineering. This study proposes a novel active control algorithm based on the neural network for real time seismic response mitigation in a single-storey building frame. The proposed algorithm is developed on the assumption that the shape of the time history of the control force is similar to the shape of the time history of the earthquake. A state-space formulation of the dynamic equilibrium equations is used to compute structural responses under seismic excitation. Synthetic ground motions compatible with the response spectra for seismic Zones III, IV, and V, as defined in IS 1893:2016, are generated and used to train the neural network. The neural network is trained offline using synthetic ground motion and target responses as input, with required control force as output. Once trained, the neural network is deployed in an online simulation to generate real-time control forces aimed at achieving predefined target reductions. The performance of the proposed algorithm is also evaluated under various time-step delays to assess its stability in real-time conditions. Results show that the proposed algorithm not only achieves but consistently exceeds the predefined target reduction levels. Time-delay analysis further confirms the stability and robustness of the control strategy under implementation constraints. This approach offers a scalable pathway toward intelligent, adaptive structural control systems for seismic risk mitigation.</p>